HORN: STROMATEOIDEI 



Table 165. Extended. 



627 



related to the stromateids than to the nomeids as proposed by 

 Haedrich (1967). Of the six families recognized by Haedrich 

 and Horn (1972), only the Amarsipidae, Nomeidae and Stro- 

 mateidae appear to be monophyletic based on the generic re- 

 lationships expressed in the cladogram. The Centrolophidae, on 

 the other hand, lacks a synapomorphy and there is no indication 

 that Tetragonuridae and Ariommidae should be considered dis- 

 tinct families. Additional characters, however, should be ex- 

 amined before a change in classification is proposed. 



The Scorpididae and Girellidae were part of a trichotomy at 

 the base of the tree and together formed the plesiomorphic out- 

 group cluster in the analysis. 



Character diagnosis showed that reversals in character trans- 

 formation occurred most frequently (> 3 taxa or stems/character 

 state) for scale type (no. 14) and juvenile pigmentation (no. 26). 

 In the same diagnosis, character convergences occurred most 

 often for the supramaxillary bone (no. 10), opercular scalation 

 (no. 15), number of branchiostegal rays (no. 18), number of 

 dorsal fins (no. 21)andjuvenileassociations(no. 27). Atnarsipus 

 was involved in all five of these cases of apparent convergence, 

 an indication of its uncertain phylogenetic position. 



The tree remains incompletely resolved with three polychot- 

 omies (Fig. 335). In addition to the trichotomy at the base of 

 the tree, the other two nodes with multiple branches involve 

 centrolophid genera. Lack of full resolution in this region of the 

 tree indicates that further work is needed to clarify the inter- 

 generic relationships of the Centrolophidae. Extending the anal- 

 ysis to the species level would provide greater resolution. 



Strictly larval or juvenile characters have contributed little to 



the broad understanding of stromaleoid intergeneric relation- 

 ships as perceived by Haedrich (1967) or as analyzed in the 

 present study. Elimination of the two juvenile characters (pig- 

 mentation and associations) from the present analysis resulted 

 in a tree virtually identical to that with them included (Fig. 335). 

 The study of the early life history stages of pelagic stromateoids 

 by Ahlstrom et al. (1976), however, is a major contribution to 

 stromateoid systematics especially in developing an approach 

 that can potentially expand to all taxa in the suborder. Their 

 use of ontogenetic characters was important at the species level 

 and particularly valuable in distinguishing the species and gen- 

 era of nomeids. 



Characters employed by Ahlstrom et al. (1976) that hold 

 promise for resolving relationships among stromateoids in gen- 

 eral include ( 1 ) sequence of fin formation, (2) arrangement of 

 anal fin pterygiophores in relation to haemal spines, (3) head 

 armature and (4) pigmentation patterns. The caudal fin complex, 

 while not representing a strictly ontogenetic suite of features, 

 also appears likely to provide characters if a full spectrum of 

 cleared and stained larvae are carefully examined. Finally, the 

 various types of associations juvenile stromateoids hold with 

 floating objects may be more specific than generally thought and 

 could become a rich source of characters. 



Relationships of the Stromateoidei to 

 other groups 



Haedrich ( 1 967) in his review of stromateoid systematics pro- 

 posed that the group arose from within a relatively undiffer- 

 entiated assemblage of perciform families including the Arri- 



